人脸生成（Face Generation）¶

在该项目中，你将使用生成式对抗网络（Generative Adversarial Nets）来生成新的人脸图像。

获取数据¶

该项目将使用以下数据集：

MNIST
CelebA

由于 CelebA 数据集比较复杂，而且这是你第一次使用 GANs。我们想让你先在 MNIST 数据集上测试你的 GANs 模型，以让你更快的评估所建立模型的性能。

如果你在使用 FloydHub, 请将 data_dir 设置为 "/input" 并使用 FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe".

data_dir = '/data'
!pip install matplotlib==2.0.2
# FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe"
#data_dir = '/input'


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper

helper.download_extract('mnist', data_dir)
helper.download_extract('celeba', data_dir)

Collecting matplotlib==2.0.2
  Downloading https://files.pythonhosted.org/packages/60/d4/6b6d8a7a6bc69a1602ab372f6fc6e88ef88a8a96398a1a25edbac636295b/matplotlib-2.0.2-cp36-cp36m-manylinux1_x86_64.whl (14.6MB)
    100% |████████████████████████████████| 14.6MB 46kB/s  eta 0:00:01   27% |████████▊                       | 4.0MB 23.0MB/s eta 0:00:01
Requirement already satisfied: pyparsing!=2.0.0,!=2.0.4,!=2.1.2,!=2.1.6,>=1.5.6 in /opt/conda/lib/python3.6/site-packages (from matplotlib==2.0.2)
Requirement already satisfied: pytz in /opt/conda/lib/python3.6/site-packages (from matplotlib==2.0.2)
Requirement already satisfied: cycler>=0.10 in /opt/conda/lib/python3.6/site-packages/cycler-0.10.0-py3.6.egg (from matplotlib==2.0.2)
Requirement already satisfied: numpy>=1.7.1 in /opt/conda/lib/python3.6/site-packages (from matplotlib==2.0.2)
Requirement already satisfied: six>=1.10 in /opt/conda/lib/python3.6/site-packages (from matplotlib==2.0.2)
Requirement already satisfied: python-dateutil in /opt/conda/lib/python3.6/site-packages (from matplotlib==2.0.2)
Installing collected packages: matplotlib
  Found existing installation: matplotlib 2.1.0
    Uninstalling matplotlib-2.1.0:
      Successfully uninstalled matplotlib-2.1.0
Successfully installed matplotlib-2.0.2
You are using pip version 9.0.1, however version 18.0 is available.
You should consider upgrading via the 'pip install --upgrade pip' command.
Found mnist Data
Found celeba Data

探索数据（Explore the Data）¶

MNIST¶

MNIST 是一个手写数字的图像数据集。你可以更改 show_n_images 探索此数据集。

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
import os
from glob import glob
from matplotlib import pyplot

mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L')
pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray')

<matplotlib.image.AxesImage at 0x7f76dfb2c588>

CelebA¶

CelebFaces Attributes Dataset (CelebA) 是一个包含 20 多万张名人图片及相关图片说明的数据集。你将用此数据集生成人脸，不会用不到相关说明。你可以更改 show_n_images 探索此数据集。

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 28, 28, 'RGB')
pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB'))

<matplotlib.image.AxesImage at 0x7f76dfa202e8>

预处理数据（Preprocess the Data）¶

由于该项目的重点是建立 GANs 模型，我们将为你预处理数据。

经过数据预处理，MNIST 和 CelebA 数据集的值在 28×28 维度图像的 [-0.5, 0.5] 范围内。CelebA 数据集中的图像裁剪了非脸部的图像部分，然后调整到 28x28 维度。

MNIST 数据集中的图像是单通道的黑白图像，CelebA 数据集中的图像是三通道的 RGB 彩色图像。

建立神经网络（Build the Neural Network）¶

你将通过部署以下函数来建立 GANs 的主要组成部分:

model_inputs
discriminator
generator
model_loss
model_opt
train

检查 TensorFlow 版本并获取 GPU 型号¶

检查你是否使用正确的 TensorFlow 版本，并获取 GPU 型号

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer.  You are using {}'.format(tf.__version__)
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))

TensorFlow Version: 1.3.0
Default GPU Device: /gpu:0

输入（Input）¶

部署 model_inputs 函数以创建用于神经网络的占位符 (TF Placeholders)。请创建以下占位符：

输入图像占位符: 使用 image_width，image_height 和 image_channels 设置为 rank 4。
输入 Z 占位符: 设置为 rank 2，并命名为 z_dim。
学习速率占位符: 设置为 rank 0。

返回占位符元组的形状为 (tensor of real input images, tensor of z data, learning rate)。

import problem_unittests as tests

def model_inputs(image_width, image_height, image_channels, z_dim):
    """
    Create the model inputs
    :param image_width: The input image width
    :param image_height: The input image height
    :param image_channels: The number of image channels
    :param z_dim: The dimension of Z
    :return: Tuple of (tensor of real input images, tensor of z data, learning rate)
    """
    # TODO: Implement Function
    input_real = tf.placeholder(tf.float32, (None, image_width, image_height, image_channels))
    input_z = tf.placeholder(tf.float32, (None, z_dim), name = 'z_dim')
    learning_rate = tf.placeholder(tf.float32,(None))
    return input_real, input_z, learning_rate


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_inputs(model_inputs)

ERROR:tensorflow:==================================
Object was never used (type <class 'tensorflow.python.framework.ops.Operation'>):
<tf.Operation 'assert_rank_2/Assert/Assert' type=Assert>
If you want to mark it as used call its "mark_used()" method.
It was originally created here:
['File "/opt/conda/lib/python3.6/runpy.py", line 193, in _run_module_as_main\n    "__main__", mod_spec)', 'File "/opt/conda/lib/python3.6/runpy.py", line 85, in _run_code\n    exec(code, run_globals)', 'File "/opt/conda/lib/python3.6/site-packages/ipykernel_launcher.py", line 16, in <module>\n    app.launch_new_instance()', 'File "/opt/conda/lib/python3.6/site-packages/traitlets/config/application.py", line 658, in launch_instance\n    app.start()', 'File "/opt/conda/lib/python3.6/site-packages/ipykernel/kernelapp.py", line 478, in start\n    self.io_loop.start()', 'File "/opt/conda/lib/python3.6/site-packages/zmq/eventloop/ioloop.py", line 177, in start\n    super(ZMQIOLoop, self).start()', 'File "/opt/conda/lib/python3.6/site-packages/tornado/ioloop.py", line 888, in start\n    handler_func(fd_obj, events)', 'File "/opt/conda/lib/python3.6/site-packages/tornado/stack_context.py", line 277, in null_wrapper\n    return fn(*args, **kwargs)', 'File "/opt/conda/lib/python3.6/site-packages/zmq/eventloop/zmqstream.py", line 440, in _handle_events\n    self._handle_recv()', 'File "/opt/conda/lib/python3.6/site-packages/zmq/eventloop/zmqstream.py", line 472, in _handle_recv\n    self._run_callback(callback, msg)', 'File "/opt/conda/lib/python3.6/site-packages/zmq/eventloop/zmqstream.py", line 414, in _run_callback\n    callback(*args, **kwargs)', 'File "/opt/conda/lib/python3.6/site-packages/tornado/stack_context.py", line 277, in null_wrapper\n    return fn(*args, **kwargs)', 'File "/opt/conda/lib/python3.6/site-packages/ipykernel/kernelbase.py", line 281, in dispatcher\n    return self.dispatch_shell(stream, msg)', 'File "/opt/conda/lib/python3.6/site-packages/ipykernel/kernelbase.py", line 232, in dispatch_shell\n    handler(stream, idents, msg)', 'File "/opt/conda/lib/python3.6/site-packages/ipykernel/kernelbase.py", line 397, in execute_request\n    user_expressions, allow_stdin)', 'File "/opt/conda/lib/python3.6/site-packages/ipykernel/ipkernel.py", line 208, in do_execute\n    res = shell.run_cell(code, store_history=store_history, silent=silent)', 'File "/opt/conda/lib/python3.6/site-packages/ipykernel/zmqshell.py", line 533, in run_cell\n    return super(ZMQInteractiveShell, self).run_cell(*args, **kwargs)', 'File "/opt/conda/lib/python3.6/site-packages/IPython/core/interactiveshell.py", line 2728, in run_cell\n    interactivity=interactivity, compiler=compiler, result=result)', 'File "/opt/conda/lib/python3.6/site-packages/IPython/core/interactiveshell.py", line 2856, in run_ast_nodes\n    if self.run_code(code, result):', 'File "/opt/conda/lib/python3.6/site-packages/IPython/core/interactiveshell.py", line 2910, in run_code\n    exec(code_obj, self.user_global_ns, self.user_ns)', 'File "<ipython-input-5-6a25c9df9163>", line 22, in <module>\n    tests.test_model_inputs(model_inputs)', 'File "/home/workspace/problem_unittests.py", line 12, in func_wrapper\n    result = func(*args)', 'File "/home/workspace/problem_unittests.py", line 68, in test_model_inputs\n    _check_input(learn_rate, [], \'Learning Rate\')', 'File "/home/workspace/problem_unittests.py", line 34, in _check_input\n    _assert_tensor_shape(tensor, shape, \'Real Input\')', 'File "/home/workspace/problem_unittests.py", line 20, in _assert_tensor_shape\n    assert tf.assert_rank(tensor, len(shape), message=\'{} has wrong rank\'.format(display_name))', 'File "/opt/conda/lib/python3.6/site-packages/tensorflow/python/ops/check_ops.py", line 617, in assert_rank\n    dynamic_condition, data, summarize)', 'File "/opt/conda/lib/python3.6/site-packages/tensorflow/python/ops/check_ops.py", line 571, in _assert_rank_condition\n    return control_flow_ops.Assert(condition, data, summarize=summarize)', 'File "/opt/conda/lib/python3.6/site-packages/tensorflow/python/util/tf_should_use.py", line 175, in wrapped\n    return _add_should_use_warning(fn(*args, **kwargs))', 'File "/opt/conda/lib/python3.6/site-packages/tensorflow/python/util/tf_should_use.py", line 144, in _add_should_use_warning\n    wrapped = TFShouldUseWarningWrapper(x)', 'File "/opt/conda/lib/python3.6/site-packages/tensorflow/python/util/tf_should_use.py", line 101, in __init__\n    stack = [s.strip() for s in traceback.format_stack()]']
==================================
Tests Passed

辨别器（Discriminator）¶

部署 discriminator 函数创建辨别器神经网络以辨别 images。该函数应能够重复使用神经网络中的各种变量。在 tf.variable_scope 中使用 "discriminator" 的变量空间名来重复使用该函数中的变量。

该函数应返回形如 (tensor output of the discriminator, tensor logits of the discriminator) 的元组。

def discriminator(images, reuse=False, alpha = 0.2):
    """
    Create the discriminator network
    :param image: Tensor of input image(s)
    :param reuse: Boolean if the weights should be reused
    :return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator)
    """
    # TODO: Implement Function
    with tf.variable_scope('discriminator', reuse=reuse):
        # Input layer is 28x28x3
        # to 14*14
        x_1 = tf.layers.conv2d(images, 64, 5, strides = 2, padding = 'same')
        x_1 = tf.maximum(alpha*x_1, x_1)
        
        # to 7*7
        x_2 = tf.layers.conv2d(x_1, 128, 5, strides = 2, padding = 'same')
        x_2 = tf.layers.batch_normalization(x_2, training = True)
        x_2 = tf.maximum(alpha*x_2, x_2)
        
        # to 4*4
        x_3 = tf.layers.conv2d(x_2, 256, 5, strides = 2, padding = 'same')
        x_3 = tf.layers.batch_normalization(x_3, training = True)
        x_3 = tf.maximum(alpha*x_3, x_3)
        
        #x_4 = tf.layers.average_pooling2d(x_3, pool_size = 4, strides = 1, padding = 'valid')
        x_4 = tf.reshape(x_3, (-1,256*4*4))
        
        logits = tf.layers.dense(x_4,1)
        out = tf.nn.sigmoid(logits)
        
        return out, logits

"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_discriminator(discriminator, tf)

Tests Passed

生成器（Generator）¶

部署 generator 函数以使用 z 生成图像。该函数应能够重复使用神经网络中的各种变量。在 tf.variable_scope 中使用 "generator" 的变量空间名来重复使用该函数中的变量。

该函数应返回所生成的 28 x 28 x out_channel_dim 维度图像。

def generator(z, out_channel_dim, is_train=True, alpha = 0.2):
    """
    Create the generator network
    :param z: Input z
    :param out_channel_dim: The number of channels in the output image
    :param is_train: Boolean if generator is being used for training
    :return: The tensor output of the generator
    """
    # TODO: Implement Function
    with tf.variable_scope( "generator", reuse = not is_train):
        # First fully connected layer
        x_full = tf.layers.dense(z, 4*4*512)
        x_0 = tf.reshape(x_full, (-1,4,4,512))
        x_0 = tf.layers.batch_normalization(x_0, training = is_train)
        x_0 = tf.maximum(alpha*x_0, x_0)
        
        # 7*7
        x_1 = tf.layers.conv2d_transpose(x_0, 256, 4, strides = 1, padding = 'valid')
        x_1 = tf.layers.batch_normalization(x_1, training = is_train)
        x_1 = tf.maximum(alpha* x_1, x_1)
        
        # 14*14
        x_2 = tf.layers.conv2d_transpose(x_1, 128, 5, strides = 2, padding = 'same')
        x_2 = tf.layers.batch_normalization(x_2, training = is_train)
        x_2 = tf.maximum(alpha*x_2, x_2)
        
        #28*28
        logits = tf.layers.conv2d_transpose(x_2, out_channel_dim, 5, strides = 2, padding = 'same')
        
        out = tf.tanh(logits)
        return out


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_generator(generator, tf)

Tests Passed

损失函数（Loss）¶

部署 model_loss 函数训练并计算 GANs 的损失。该函数应返回形如 (discriminator loss, generator loss) 的元组。

使用你已实现的函数：

discriminator(images, reuse=False)
generator(z, out_channel_dim, is_train=True)

def model_loss(input_real, input_z, out_channel_dim, learning_rate = 0.0002, smooth = 0.1):
    """
    Get the loss for the discriminator and generator
    :param input_real: Images from the real dataset
    :param input_z: Z input
    :param out_channel_dim: The number of channels in the output image
    :return: A tuple of (discriminator loss, generator loss)
    """
    # TODO: Implement Function
    d_output_real, d_logit_real = discriminator(input_real)
    g_output = generator(input_z, out_channel_dim)
    d_output_fake, d_logit_fake = discriminator(g_output, reuse = True)
    
    d_loss_real = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits = d_logit_real, labels = tf.ones_like(d_output_real)*(1-smooth)))
    d_loss_fake = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits = d_logit_fake, labels = tf.zeros_like(d_output_fake)))
    g_loss = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits = d_logit_fake, labels = tf.ones_like(d_output_fake)))
    
    d_loss = d_loss_real + d_loss_fake
    return d_loss, g_loss


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_loss(model_loss)

Tests Passed

优化（Optimization）¶

部署 model_opt 函数实现对 GANs 的优化。使用 tf.trainable_variables 获取可训练的所有变量。通过变量空间名 discriminator 和 generator 来过滤变量。该函数应返回形如 (discriminator training operation, generator training operation) 的元组。

def model_opt(d_loss, g_loss, learning_rate, beta1):
    """
    Get optimization operations
    :param d_loss: Discriminator loss Tensor
    :param g_loss: Generator loss Tensor
    :param learning_rate: Learning Rate Placeholder
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :return: A tuple of (discriminator training operation, generator training operation)
    """
    # TODO: Implement Function
    t_vars = tf.trainable_variables()
    g_vars = [var for var in t_vars if var.name.startswith('generator')]
    d_vars = [var for var in t_vars if var.name.startswith('discriminator')]
    
    with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
        d_opt = tf.train.AdamOptimizer(learning_rate,beta1 = beta1).minimize(d_loss, var_list = d_vars)
        g_opt = tf.train.AdamOptimizer(learning_rate,beta1 = beta1).minimize(g_loss, var_list = g_vars)
        return d_opt, g_opt


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_opt(model_opt, tf)

Tests Passed

训练神经网络（Neural Network Training）¶

输出显示¶

使用该函数可以显示生成器 (Generator) 在训练过程中的当前输出，这会帮你评估 GANs 模型的训练程度。

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode):
    """
    Show example output for the generator
    :param sess: TensorFlow session
    :param n_images: Number of Images to display
    :param input_z: Input Z Tensor
    :param out_channel_dim: The number of channels in the output image
    :param image_mode: The mode to use for images ("RGB" or "L")
    """
    cmap = None if image_mode == 'RGB' else 'gray'
    z_dim = input_z.get_shape().as_list()[-1]
    example_z = np.random.uniform(-1, 1, size=[n_images, z_dim])

    samples = sess.run(
        generator(input_z, out_channel_dim, False),
        feed_dict={input_z: example_z})

    images_grid = helper.images_square_grid(samples, image_mode)
    pyplot.imshow(images_grid, cmap=cmap)
    pyplot.show()

训练¶

部署 train 函数以建立并训练 GANs 模型。记得使用以下你已完成的函数：

model_inputs(image_width, image_height, image_channels, z_dim)
model_loss(input_real, input_z, out_channel_dim)
model_opt(d_loss, g_loss, learning_rate, beta1)

使用 show_generator_output 函数显示 generator 在训练过程中的输出。

注意：在每个批次 (batch) 中运行 show_generator_output 函数会显著增加训练时间与该 notebook 的体积。推荐每 100 批次输出一次 generator 的输出。

def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode):
    """
    Train the GAN
    :param epoch_count: Number of epochs
    :param batch_size: Batch Size
    :param z_dim: Z dimension
    :param learning_rate: Learning Rate
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :param get_batches: Function to get batches
    :param data_shape: Shape of the data
    :param data_image_mode: The image mode to use for images ("RGB" or "L")
    """
    # Build Model
    input_real, input_z, lr = model_inputs(data_shape[1],data_shape[2],data_shape[3], z_dim)
    d_loss,g_loss = model_loss(input_real, input_z, data_shape[3])
    d_opt, g_opt = model_opt(d_loss, g_loss, lr, beta1)
    
    with tf.Session() as sess:
        sess.run(tf.global_variables_initializer())
        print_loss = 0
        show_generator = 200
        for epoch_i in range(epoch_count):
            for batch_images in get_batches(batch_size):
                print_loss += 1
                z = np.random.uniform(-1,1,size =(batch_size, z_dim))
                # update weights
                sess.run(d_opt, feed_dict = {input_real: batch_images,
                                                     input_z: z,
                                                     lr: learning_rate})
                sess.run(g_opt, feed_dict = {input_z: z,
                                             lr:learning_rate,
                                             input_real: batch_images})
                
                # print loss every 10 batches
                if print_loss % 20 == 0:
                    train_d_loss = d_loss.eval({input_real: batch_images, input_z: z})
                    train_g_loss = g_loss.eval({input_z: z})
                    print('epoch:{} batch: {}, genarator loss:{:4f}, discriminator loss:{:4f}'.
                          format(epoch_i, print_loss, train_g_loss, train_d_loss))
                
                # show generator output
                if print_loss % show_generator == 0:
                    show_generator_output(sess, batch_size, input_z, data_shape[3], data_image_mode)

MNIST¶

在 MNIST 上测试你的 GANs 模型。经过 2 次迭代，GANs 应该能够生成类似手写数字的图像。确保生成器 (generator) 低于辨别器 (discriminator) 的损失，或接近 0。

batch_size = 16
z_dim = 100
learning_rate = 0.0002
beta1 = 0.5


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 2

mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches,
          mnist_dataset.shape, mnist_dataset.image_mode)

epoch:0 batch: 10, genarator loss:2.040328, discriminator loss:0.542080
epoch:0 batch: 20, genarator loss:3.452676, discriminator loss:0.376089
epoch:0 batch: 30, genarator loss:2.418956, discriminator loss:0.441421
epoch:0 batch: 40, genarator loss:17.602821, discriminator loss:0.433780
epoch:0 batch: 50, genarator loss:10.225800, discriminator loss:0.712103
epoch:0 batch: 60, genarator loss:10.630457, discriminator loss:1.134504
epoch:0 batch: 70, genarator loss:0.575330, discriminator loss:1.276171
epoch:0 batch: 80, genarator loss:7.430463, discriminator loss:0.590977
epoch:0 batch: 90, genarator loss:0.965443, discriminator loss:0.995948
epoch:0 batch: 100, genarator loss:2.900015, discriminator loss:0.472830

epoch:0 batch: 110, genarator loss:3.905557, discriminator loss:0.410138
epoch:0 batch: 120, genarator loss:1.596554, discriminator loss:0.635088
epoch:0 batch: 130, genarator loss:3.027503, discriminator loss:0.408492
epoch:0 batch: 140, genarator loss:2.789096, discriminator loss:0.451580
epoch:0 batch: 150, genarator loss:3.212376, discriminator loss:0.396580
epoch:0 batch: 160, genarator loss:4.119608, discriminator loss:0.371187
epoch:0 batch: 170, genarator loss:3.473252, discriminator loss:0.394308
epoch:0 batch: 180, genarator loss:2.912415, discriminator loss:0.431691
epoch:0 batch: 190, genarator loss:3.843958, discriminator loss:0.367091
epoch:0 batch: 200, genarator loss:2.684896, discriminator loss:0.443534

epoch:0 batch: 210, genarator loss:2.380062, discriminator loss:0.474473
epoch:0 batch: 220, genarator loss:3.359319, discriminator loss:0.400293
epoch:0 batch: 230, genarator loss:3.972869, discriminator loss:0.378608
epoch:0 batch: 240, genarator loss:2.549768, discriminator loss:0.485034
epoch:0 batch: 250, genarator loss:1.920704, discriminator loss:0.653034
epoch:0 batch: 260, genarator loss:2.142667, discriminator loss:1.353784
epoch:0 batch: 270, genarator loss:2.176462, discriminator loss:0.545984
epoch:0 batch: 280, genarator loss:0.934039, discriminator loss:0.931761
epoch:0 batch: 290, genarator loss:3.211729, discriminator loss:0.450786
epoch:0 batch: 300, genarator loss:2.369772, discriminator loss:0.488948

epoch:0 batch: 310, genarator loss:3.485872, discriminator loss:0.402823
epoch:0 batch: 320, genarator loss:2.734115, discriminator loss:0.465706
epoch:0 batch: 330, genarator loss:4.579434, discriminator loss:0.491517
epoch:0 batch: 340, genarator loss:2.438935, discriminator loss:0.449288
epoch:0 batch: 350, genarator loss:4.432307, discriminator loss:0.502139
epoch:0 batch: 360, genarator loss:3.153699, discriminator loss:0.386491
epoch:0 batch: 370, genarator loss:3.324399, discriminator loss:0.419091
epoch:0 batch: 380, genarator loss:3.752177, discriminator loss:0.439421
epoch:0 batch: 390, genarator loss:2.898462, discriminator loss:0.446318
epoch:0 batch: 400, genarator loss:4.303602, discriminator loss:0.390507

epoch:0 batch: 410, genarator loss:4.139005, discriminator loss:0.405757
epoch:0 batch: 420, genarator loss:3.317191, discriminator loss:0.384216
epoch:0 batch: 430, genarator loss:2.813351, discriminator loss:0.409549
epoch:0 batch: 440, genarator loss:2.446928, discriminator loss:0.443006
epoch:0 batch: 450, genarator loss:0.079324, discriminator loss:3.794870
epoch:0 batch: 460, genarator loss:1.973103, discriminator loss:0.573760
epoch:0 batch: 470, genarator loss:1.141652, discriminator loss:0.942307
epoch:0 batch: 480, genarator loss:3.954948, discriminator loss:1.307546
epoch:0 batch: 490, genarator loss:0.221785, discriminator loss:2.377927
epoch:0 batch: 500, genarator loss:4.316993, discriminator loss:0.852278

epoch:0 batch: 510, genarator loss:1.157313, discriminator loss:0.887464
epoch:0 batch: 520, genarator loss:3.409136, discriminator loss:0.425414
epoch:0 batch: 530, genarator loss:3.563782, discriminator loss:0.689708
epoch:0 batch: 540, genarator loss:0.281096, discriminator loss:1.870297
epoch:0 batch: 550, genarator loss:1.446675, discriminator loss:0.772441
epoch:0 batch: 560, genarator loss:1.586989, discriminator loss:0.695982
epoch:0 batch: 570, genarator loss:2.385432, discriminator loss:0.464247
epoch:0 batch: 580, genarator loss:1.942888, discriminator loss:0.534875
epoch:0 batch: 590, genarator loss:4.030513, discriminator loss:0.358684
epoch:0 batch: 600, genarator loss:1.395116, discriminator loss:0.706214

epoch:0 batch: 610, genarator loss:1.132258, discriminator loss:0.798133
epoch:0 batch: 620, genarator loss:1.641075, discriminator loss:0.632637
epoch:0 batch: 630, genarator loss:1.433766, discriminator loss:0.651574
epoch:0 batch: 640, genarator loss:3.003560, discriminator loss:0.480912
epoch:0 batch: 650, genarator loss:1.136741, discriminator loss:0.799359
epoch:0 batch: 660, genarator loss:1.392403, discriminator loss:0.716975
epoch:0 batch: 670, genarator loss:0.734838, discriminator loss:1.107418
epoch:0 batch: 680, genarator loss:2.450330, discriminator loss:0.484726
epoch:0 batch: 690, genarator loss:1.959856, discriminator loss:0.505301
epoch:0 batch: 700, genarator loss:2.938405, discriminator loss:0.405105

epoch:0 batch: 710, genarator loss:2.540901, discriminator loss:0.478950
epoch:0 batch: 720, genarator loss:2.384749, discriminator loss:0.439710
epoch:0 batch: 730, genarator loss:2.407166, discriminator loss:0.500783
epoch:0 batch: 740, genarator loss:2.381240, discriminator loss:0.499645
epoch:0 batch: 750, genarator loss:3.668548, discriminator loss:0.391520
epoch:0 batch: 760, genarator loss:2.235296, discriminator loss:0.484967
epoch:0 batch: 770, genarator loss:2.733843, discriminator loss:0.435325
epoch:0 batch: 780, genarator loss:3.203180, discriminator loss:0.408801
epoch:0 batch: 790, genarator loss:2.527978, discriminator loss:0.447681
epoch:0 batch: 800, genarator loss:2.653310, discriminator loss:0.426497

epoch:0 batch: 810, genarator loss:1.967795, discriminator loss:0.519500
epoch:0 batch: 820, genarator loss:2.309624, discriminator loss:0.458722
epoch:0 batch: 830, genarator loss:1.004939, discriminator loss:0.865088
epoch:0 batch: 840, genarator loss:1.700193, discriminator loss:0.575222
epoch:0 batch: 850, genarator loss:1.512322, discriminator loss:0.628559
epoch:0 batch: 860, genarator loss:1.621896, discriminator loss:0.569687
epoch:0 batch: 870, genarator loss:1.767615, discriminator loss:0.566540
epoch:0 batch: 880, genarator loss:1.009246, discriminator loss:0.945293
epoch:0 batch: 890, genarator loss:1.649165, discriminator loss:0.567871
epoch:0 batch: 900, genarator loss:3.064413, discriminator loss:0.393825

epoch:0 batch: 910, genarator loss:1.360594, discriminator loss:0.708579
epoch:0 batch: 920, genarator loss:1.850842, discriminator loss:0.631701
epoch:0 batch: 930, genarator loss:2.825734, discriminator loss:0.420743
epoch:0 batch: 940, genarator loss:2.456910, discriminator loss:0.465818
epoch:0 batch: 950, genarator loss:1.458563, discriminator loss:1.010305
epoch:0 batch: 960, genarator loss:1.932614, discriminator loss:0.540064
epoch:0 batch: 970, genarator loss:0.915047, discriminator loss:0.963529
epoch:0 batch: 980, genarator loss:1.798701, discriminator loss:0.542726
epoch:0 batch: 990, genarator loss:2.441283, discriminator loss:0.452009
epoch:0 batch: 1000, genarator loss:2.152303, discriminator loss:0.475002

CelebA¶

在 CelebA 上运行你的 GANs 模型。在一般的GPU上运行每次迭代大约需要 20 分钟。你可以运行整个迭代，或者当 GANs 开始产生真实人脸图像时停止它。

batch_size = 8
z_dim = 100
learning_rate = 0.0002
beta1 = 0.5


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 1

celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches,
          celeba_dataset.shape, celeba_dataset.image_mode)

epoch:0 batch: 20, genarator loss:2.397509, discriminator loss:0.471767
epoch:0 batch: 40, genarator loss:2.389175, discriminator loss:0.533117
epoch:0 batch: 60, genarator loss:2.985783, discriminator loss:0.549669
epoch:0 batch: 80, genarator loss:10.747932, discriminator loss:0.623940
epoch:0 batch: 100, genarator loss:5.490516, discriminator loss:0.672262
epoch:0 batch: 120, genarator loss:2.661679, discriminator loss:0.479137
epoch:0 batch: 140, genarator loss:3.467271, discriminator loss:0.442595
epoch:0 batch: 160, genarator loss:5.467538, discriminator loss:0.404987
epoch:0 batch: 180, genarator loss:2.570015, discriminator loss:0.436779
epoch:0 batch: 200, genarator loss:1.589005, discriminator loss:0.618640

epoch:0 batch: 220, genarator loss:3.774475, discriminator loss:0.387900
epoch:0 batch: 240, genarator loss:3.569595, discriminator loss:0.376694
epoch:0 batch: 260, genarator loss:3.303669, discriminator loss:0.405830
epoch:0 batch: 280, genarator loss:4.370145, discriminator loss:0.390278
epoch:0 batch: 300, genarator loss:2.409434, discriminator loss:0.437721
epoch:0 batch: 320, genarator loss:1.327529, discriminator loss:0.679740
epoch:0 batch: 340, genarator loss:8.907648, discriminator loss:0.341813
epoch:0 batch: 360, genarator loss:6.808239, discriminator loss:0.415230
epoch:0 batch: 380, genarator loss:4.058443, discriminator loss:0.371019
epoch:0 batch: 400, genarator loss:5.949522, discriminator loss:0.428813

epoch:0 batch: 420, genarator loss:6.757037, discriminator loss:0.560681
epoch:0 batch: 440, genarator loss:3.859785, discriminator loss:0.382167
epoch:0 batch: 460, genarator loss:3.635761, discriminator loss:0.394882
epoch:0 batch: 480, genarator loss:7.333775, discriminator loss:0.354470
epoch:0 batch: 500, genarator loss:2.325486, discriminator loss:0.484315
epoch:0 batch: 520, genarator loss:0.238382, discriminator loss:2.238207
epoch:0 batch: 540, genarator loss:0.471795, discriminator loss:1.673427
epoch:0 batch: 560, genarator loss:2.855639, discriminator loss:0.414017
epoch:0 batch: 580, genarator loss:1.703179, discriminator loss:0.612272
epoch:0 batch: 600, genarator loss:2.384623, discriminator loss:0.558816

epoch:0 batch: 620, genarator loss:2.384628, discriminator loss:0.551793
epoch:0 batch: 640, genarator loss:1.524964, discriminator loss:0.798444
epoch:0 batch: 660, genarator loss:3.442170, discriminator loss:0.406504
epoch:0 batch: 680, genarator loss:3.036144, discriminator loss:0.422467
epoch:0 batch: 700, genarator loss:4.172835, discriminator loss:0.458799
epoch:0 batch: 720, genarator loss:0.450846, discriminator loss:1.513631
epoch:0 batch: 740, genarator loss:1.573265, discriminator loss:0.676565
epoch:0 batch: 760, genarator loss:2.799005, discriminator loss:0.509965
epoch:0 batch: 780, genarator loss:1.120569, discriminator loss:0.800954
epoch:0 batch: 800, genarator loss:4.421659, discriminator loss:0.519817

epoch:0 batch: 820, genarator loss:3.632126, discriminator loss:0.437697
epoch:0 batch: 840, genarator loss:0.484206, discriminator loss:1.703066
epoch:0 batch: 860, genarator loss:2.680720, discriminator loss:0.447329
epoch:0 batch: 880, genarator loss:0.776103, discriminator loss:1.016839
epoch:0 batch: 900, genarator loss:5.346769, discriminator loss:0.413979
epoch:0 batch: 920, genarator loss:2.613636, discriminator loss:0.795822
epoch:0 batch: 940, genarator loss:5.418452, discriminator loss:0.348929
epoch:0 batch: 960, genarator loss:6.650192, discriminator loss:0.835514
epoch:0 batch: 980, genarator loss:1.144597, discriminator loss:0.747543
epoch:0 batch: 1000, genarator loss:1.697827, discriminator loss:0.693509

epoch:0 batch: 1020, genarator loss:3.816097, discriminator loss:0.383830
epoch:0 batch: 1040, genarator loss:4.357079, discriminator loss:0.419309
epoch:0 batch: 1060, genarator loss:3.877601, discriminator loss:0.386302
epoch:0 batch: 1080, genarator loss:1.084211, discriminator loss:0.920076
epoch:0 batch: 1100, genarator loss:5.393482, discriminator loss:0.360638
epoch:0 batch: 1120, genarator loss:3.192264, discriminator loss:0.383260
epoch:0 batch: 1140, genarator loss:4.091304, discriminator loss:0.382516
epoch:0 batch: 1160, genarator loss:5.758044, discriminator loss:0.380051
epoch:0 batch: 1180, genarator loss:0.318344, discriminator loss:1.825161
epoch:0 batch: 1200, genarator loss:8.465484, discriminator loss:0.331630

epoch:0 batch: 1220, genarator loss:8.469044, discriminator loss:0.521228
epoch:0 batch: 1240, genarator loss:1.280100, discriminator loss:0.771811
epoch:0 batch: 1260, genarator loss:0.799917, discriminator loss:0.975050
epoch:0 batch: 1280, genarator loss:7.026223, discriminator loss:0.347053
epoch:0 batch: 1300, genarator loss:2.707197, discriminator loss:0.449075
epoch:0 batch: 1320, genarator loss:6.982719, discriminator loss:0.347302
epoch:0 batch: 1340, genarator loss:2.986230, discriminator loss:0.408818
epoch:0 batch: 1360, genarator loss:6.251449, discriminator loss:0.382858
epoch:0 batch: 1380, genarator loss:8.780497, discriminator loss:0.328248
epoch:0 batch: 1400, genarator loss:1.375915, discriminator loss:0.642595

epoch:0 batch: 1420, genarator loss:11.963644, discriminator loss:0.365832
epoch:0 batch: 1440, genarator loss:9.316750, discriminator loss:0.333269
epoch:0 batch: 1460, genarator loss:0.988341, discriminator loss:0.876184
epoch:0 batch: 1480, genarator loss:5.103901, discriminator loss:0.342386
epoch:0 batch: 1500, genarator loss:5.899005, discriminator loss:0.339742
epoch:0 batch: 1520, genarator loss:9.018479, discriminator loss:0.336377
epoch:0 batch: 1540, genarator loss:3.319461, discriminator loss:0.413621
epoch:0 batch: 1560, genarator loss:4.995754, discriminator loss:0.398884
epoch:0 batch: 1580, genarator loss:8.047201, discriminator loss:0.333079
epoch:0 batch: 1600, genarator loss:1.076978, discriminator loss:0.783199

epoch:0 batch: 1620, genarator loss:8.002923, discriminator loss:0.389626
epoch:0 batch: 1640, genarator loss:5.301178, discriminator loss:0.427671
epoch:0 batch: 1660, genarator loss:0.353876, discriminator loss:1.741498
epoch:0 batch: 1680, genarator loss:0.507604, discriminator loss:1.362984
epoch:0 batch: 1700, genarator loss:9.051210, discriminator loss:0.347569
epoch:0 batch: 1720, genarator loss:1.010123, discriminator loss:0.845880
epoch:0 batch: 1740, genarator loss:8.006205, discriminator loss:0.362572
epoch:0 batch: 1760, genarator loss:8.871735, discriminator loss:0.339013
epoch:0 batch: 1780, genarator loss:4.002483, discriminator loss:0.357064
epoch:0 batch: 1800, genarator loss:0.581484, discriminator loss:1.209727

epoch:0 batch: 1820, genarator loss:9.830271, discriminator loss:0.355647
epoch:0 batch: 1840, genarator loss:10.024723, discriminator loss:0.366055
epoch:0 batch: 1860, genarator loss:2.973530, discriminator loss:0.414871
epoch:0 batch: 1880, genarator loss:5.505982, discriminator loss:0.348927
epoch:0 batch: 1900, genarator loss:7.784779, discriminator loss:0.329771
epoch:0 batch: 1920, genarator loss:10.144320, discriminator loss:0.345946
epoch:0 batch: 1940, genarator loss:3.531888, discriminator loss:0.376645
epoch:0 batch: 1960, genarator loss:3.676717, discriminator loss:0.427738
epoch:0 batch: 1980, genarator loss:9.543991, discriminator loss:0.385488
epoch:0 batch: 2000, genarator loss:12.360701, discriminator loss:0.330505

提交项目¶

提交本项目前，确保运行所有 cells 后保存该文件。

保存该文件为 "dlnd_face_generation.ipynb"，并另存为 HTML 格式 "File" -> "Download as"。提交项目时请附带 "helper.py" 和 "problem_unittests.py" 文件。